FANCD2–FANCI surveys DNA and recognizes double- to single-stranded junctions

Cloning and protein purification

A codon-optimized construct of full-length Gallus gallus FANCD2 including a carboxyl-terminal double StrepII tag was synthesized (GeneArt) and cloned into pACEBac1. The ybbR protein tag comprising the sequence DSLEFIASKLA was fused to the C terminus of FANCD2 by polymerase chain reaction (PCR). A similar strategy was used for the generation of ybbR-tagged FANCI, except the FANCI gene contained a C-terminal extension with a 3C protease site and 6×His-tag before the ybbR tag (protein sequences in Supplementary Data 1). The resulting pACEBac1 vectors containing either tagged FANCD2 or tagged FANCI were transformed into EMBacY Escherichia coli competent cells for bacmid generation. The purified bacmids were transfected into Sf9 cells (Oxford Expression Technologies Ltd, Catalogue No. 600100) and the viruses passaged twice before large-scale cultures were infected using 5 ml of P2 virus in 500 ml of Sf9 cells at 1.5 × 10⁶cells per millilitre. Cells were harvested on growth arrest, typically 2 or 3 days after infection. Sf9 cells were not authenticated but were tested regularly for Mycoplasma.

Purification of ybbR-tagged FANCD2 was carried out as previously described for FANCD2 (ref. ⁹): cells were lysed by sonication in lysis buffer (100 mM HEPES (pH 7.5), 300 mM NaCl, 1 mM TCEP, 5% glycerol, EDTA-free protease inhibitor (Roche), 5 mM benzamidine hydrochloride and 100 U ml⁻¹ Benzonase). The clarified cell lysate was incubated with Strep-Tactin Sepharose high-performance resin (GE Healthcare Life Sciences) for 60 min. The loaded resin was poured into a glass column and washed twice with lysis buffer before elution with 8 mM D-desthiobiotin. The elution was then diluted to roughly 100 mM NaCl and loaded onto a HiTrap Heparin HP affinity column (GE Healthcare Life Sciences). Using a shallow NaCl gradient, ybbR-tagged FANCD2 eluted at about 500 mM NaCl. FANCD2 was concentrated and run on a Superdex 200 26/60 column (GE Healthcare Life Sciences) in 50 mM HEPES (pH 7.5), 150 mM NaCl and 1 mM TCEP. Fractions containing pure ybbR-tagged FANCD2 were pooled, concentrated to roughly 10 mg ml⁻¹ and flash frozen for storage at −80 °C or used immediately for fluorescence labelling.

Similarly, purification of ybbR-tagged FANCI was performed as previously described for FANCI⁹: clarified cell lysate produced as for FANCD2 was loaded onto a HisTrap HP column (GE Healthcare Life Sciences). Using an imidazole gradient, FANCI eluted at about 200 mM imidazole. Collected fractions containing FANCI were diluted to 100 mM NaCl and loaded onto a HiTrap Heparin HP affinity column (GE Healthcare Life Sciences). Using a shallow NaCl gradient, FANCI eluted at roughly 500 mM NaCl. FANCI was then run on a Superdex 200 26/60 column (GE Healthcare Life Sciences) in 50 mM HEPES (pH 7.5), 150 mM NaCl and 1 mM TCEP. Fractions containing FANCI were pooled and concentrated to about 10 mg ml⁻¹ and flash frozen for storage at −80 °C or used immediately for fluorescence labelling.

Human FANCD2 and FANCI, which both carry a carboxyl-terminal double StrepII tag, were synthesized and cloned into pACEBac1 (Epoch Life Science). Human FANCD2 and FANCI were expressed and purified separately as individual proteins in the same way as chicken FANCD2 and FANCI with two adjustments. Both human FANCD2 and FANCI were incubated with Strep-Tactin resin (IBA), eluted with 8 mM D-desthiobiotin and incubated with the 3 C protease (1:100 ratio of protease to the protein of interest) to cleave the carboxyl-terminal tag for 16 h at 4 °C. Following heparin affinity, human FANCD2 and FANCI were concentrated and run on a Superdex 200 16/60 column (GE Healthcare Life Sciences). Fractions containing the protein of interest were finally concentrated to about 4.5 mg ml⁻¹ and flash frozen for storage at −80 °C.

Sample purity was assessed at each step of the purification by sodium dodecyl sulfate (SDS)–polyacrylamide gel electrophoresis (PAGE) using 4–12% NuPAGE Bis-Tris gels (Thermo Fisher Scientific), and the Gel Doc XR+ system (Bio-Rad) was used for gel imaging. Throughout purification, we routinely monitored the absorbance at 260 nm and 280 nm. The A260 nm to A280 nm ratio showed that there is no substantial nucleic acid contamination in the protein purifications.

To generate the FANCD2 KR helix mutants, the respective lysines or arginines were mutated to alanine or glutamic acid by replacing the KR helix in wild-type FANCD2 with the respective gBlocks (Integrated DNA Technologies (IDT); Supplementary Table 1) using Gibson assembly. Sanger sequencing confirmed the correct substitutions. Mutant FANCD2 proteins were expressed and purified as described for wild-type FANCD2.

Biotinylated LacR protein was prepared as previously described³⁷.

Fluorescent protein labelling

To generate site-specific fluorescently labelled proteins, we used SFP synthase (gift from J. Rhodes, Medical Research Council (MRC) Laboratory of Molecular Biology (LMB)) to conjugate CoA-activated fluorophores to the carboxyl-terminal ybbR tag. We incubated ybbR-tagged full-length FANCD2 with SFP synthase and a modified Cy3 (LD555-CoA, Lumidyne Technologies) at a 1:5:5 molar ratio for 16 h at 4 °C in 50 mM HEPES (pH 7.5), 150 mM NaCl, 1 mM TCEP and 10 mM MgCl₂. Similarly, ybbR-tagged full-length FANCI was incubated with SFP synthase and a modified Cy5 (LD655-CoA, Lumidyne Technologies) using the same conditions. Labelled protein was purified from excess free dye by gel filtration using a Superdex 200 10/300 column (GE Healthcare Life Sciences) equilibrated in 50 mM HEPES (pH 7.5), 150 mM NaCl, 1 mM TCEP and 1 mM MgCl₂. Fractions containing labelled protein (either FANCD2-Cy3 or FANCI-Cy5) were pooled, concentrated to 1–2 µM and aliquoted, flash frozen in liquid N₂ and stored at −80 °C until further use. The labelling efficiency was estimated using the extinction coefficients of ybbR-tagged FANCD2 or FANCI and either Cy3 or Cy5, respectively. All labelling reactions yielded an estimated 90–95% efficiency.

DNA binding experiments

To assess the DNA-binding activity of fluorescently labelled D2–I, we carried out electrophoretic mobility shift assays (EMSAs) as previously described^9,11. Fluorescently labelled dsDNA (purchased from IDT) was prepared by incubating complementary oligonucleotides P1(3′ FAM-labelled) and P7 (Supplementary Table 1) at 95 °C for 5 min and slowly cooling down to room temperature over roughly 2 h. For EMSAs, a 20 μl reaction containing 20 nM DNA was incubated with the indicated concentration of protein in the presence of 50 mM HEPES (pH 8.0), 75 mM NaCl and 1 mM TCEP for 30 min at 22 °C. After incubation, a 5 μl aliquot was directly loaded onto a native polyacrylamide gel (6% DNA Retardation, Thermo Fisher) and run at 4 °C in 0.5× TBE buffer for 60 min. The gel was then visualized using a Typhoon Imaging System (GE Healthcare). Each binding experiment was repeated three times (Extended Data Fig. 1c).

SwitchSENSE

The interactions of D2–I^WT, D2^EEEE–I, D2^KKEE–I and D2^EERR–I with dsDNA were analysed on a DRX2 instrument (Dynamic Biosensors GmbH) using a MPC2-48-2-G1R1-S chip equilibrated with SwitchSENSE buffer (20 mM HEPES (pH 8.0), 75 mM NaCl, 1 mM TCEP) at 25 °C. Before each kinetic analysis, 30-48-cNLB, 38-cNLB, 38P-cNLB or cNLB, together with cNLA (Supplementary Table 1), were annealed to DNA strands attached to the chip surface (NLB and NLA, respectively) by flowing 500 nM oligonucleotide over the chip for 4 min in a buffer of 10 mM Tris-HCl, 40 mM NaCl, 0.05% (v/v) Tween-20, 50 μM EDTA and 50 μM EGTA. As a result, either 48 bp dsDNA nanolevers with a 30 bp ss 5′ overhang were formed or a 38 bp dsDNA nanolever with a 10 bp 3′ overhang with a phosphate at the 5′ end was formed. Proteins in a 1:3 dilution series, starting at 100 nM protein, were injected at 50 µl min⁻¹ for 5–10 min followed by dissociation in running buffer after the highest concentration for 80 min at 50 µl min⁻¹. Dynamic switching data were analysed using the supplied switchANALYSIS v.1.9.0.33 software using a 1:1 kinetic model to give values for the association rate constant k_on and the dissociation rate constant k_off, and to calculate the kinetic dissociation constant K_d = k_off/k_on.

Ubiquitination assays

To confirm that fluorescent labelling did not affect the activity of the D2–I complex, we performed FANCL-mediated monoubiquitination assays as previously described³⁸. The reaction is based on previously described ubiquitination assays^{9,11,14,38,39}: 75 nM human (hs) E1 ubiquitin activating enzyme (Boston Biochem), 0.8 μM E2 (hsUbe2Tv4), 1 μM E3 (hsFANCL^109–375), 1 μM D2–I, 5 μM dsDNA (oligos P1 and P7; Supplementary Table 1) and 20 μM His-tagged ubiquitin (Enzo Life Sciences) were used in a total volume of 10 µl with a reaction buffer of 50 mM HEPES (pH 7.5), 64 mM NaCl, 4% glycerol, 5 mM MgCl₂, 2 mM ATP and 0.5 mM DTT. The reaction was incubated at 30 °C for 90 min and samples were analysed by SDS–PAGE (Extended Data Fig. 1d).

Preparation of monoubiquitinated FANCD2

To prepare Cy3 labelled monoubiquitinated FANCD2 for single-molecule imaging, an in vitro reaction was performed by mixing 75 nM hsE1 ubiquitin activating enzyme (Boston Biochem), 0.8 µM hsUbe2Tv4 (ref. ³⁸), 3 µM hsFANCL^109–375, 1 µM Gallus gallus FANCD2_ybbR and 20 µM His-tagged ubiquitin (Enzo Life Sciences) in a reaction buffer of 50 mM HEPES (pH 7.5), 64 mM NaCl, 4% glycerol, 5 mM MgCl₂, 2 mM ATP and 0.5 mM DTT in a total of 500 µl. The reaction was incubated at 30 °C for 90 min before applying it to 50 μl of Ni–NTA agarose resin (Qiagen) pre-equilibrated in W25 buffer (20 mM HEPES (pH 7.5), 150 mM NaCl, 1 mM TCEP and 25 mM imidazole) in a 1.5 ml centrifuge tube at 4 °C for 60 min. The resin was washed twice with 100 μl of W25 buffer (20 mM HEPES (pH 7.5), 150 mM NaCl, 1 mM TCEP and 25 mM imidazole). Each wash was performed for 30 min at 4 °C under rotation. The Ni–NTA-bound ubFANCD2 was eluted with W100 buffer (20 mM HEPES (pH 7.5), 150 mM NaCl, 1 mM TCEP and 100 mM imidazole) as previously described for the purification of ubD2–I⁹. We incubated ubFANCD2_ybbR with SFP synthase and Cy3 (LD555-CoA, Lumidyne Technologies) at a 1:5:5 molar ratio for 16 h at 4 °C in 50 mM HEPES (pH 7.5), 150 mM NaCl, 1 mM TCEP and 10 mM MgCl₂. Labelled ubFANCD2 was further purified from free dye by gel filtration using a Superdex 200 10/300 column (GE Healthcare Life Sciences) equilibrated in 50 mM HEPES (pH 7.5), 150 mM NaCl, 1 mM TCEP and 1 mM MgCl₂ (Extended Data Fig. 3g). Fractions containing labelled ubFANCD2 were pooled, concentrated to 1–2 µM and aliquoted before being flash frozen on liquid N₂, and stored at −80 °C until further use. The labelling efficiency was estimated using the extinction coefficients of ybbR-tagged FANCD2 and Cy3. All labelling reactions yielded an estimated 90–95% efficiency.

DNA for single-molecule studies

Bacteriophage λ DNA (Thermo Scientific) was labelled at both ends with biotin using Klenow polymerase exo⁻ (New England Biolabs). The linear substrate (4 nM) was incubated with 100 μM dGTP, 100 μM dTTP, 80 μM biotin-14-dATP and 80 μM biotin-14-dCTP and the enzyme (0.5 U from stock of 50 U μl⁻¹) in NEB2 buffer at 37 °C for 30 min, then at 70 °C for 15 min and subsequently cooled on ice. The product was purified using Qiagen PCR clean-up kit.

To prepare λ DNA with a site-specific ssDNA gap, a λ DNA construct was treated with CRISPR–Cas9_D10A nickase (nCas9) (IDT) as described in ref. ²⁷. In brief, λ DNA and two ssDNA biotinylated oligos (cap 1 and cap 2; Supplementary Table 1) were phosphorylated and subsequently ligated together using T4 ligase at 37 °C for 1 h to form λ DNA with closed, biotinylated ends⁴⁰. To quench the reaction, T4 ligase enzyme was inactivated at 65 °C for 20 min. Next, two nCas9–RNA complexes (complex 1: nCas9 + tracrRNA + crRNA 1, complex 2: nCas9 + tracrRNA + crRNA 2; Supplementary Table 1) were incubated in Cas9 digestion buffer (50 mM Tris-HCl (pH 8), 100 mM NaCl and 10 mM MgCl₂) with the biotinylated λ DNA at 37 °C to induce two site-specific nicks. crRNA 1 was labelled with a Cy3 fluorescent dye to generate the fiduciary static nCas9 marker. For D2–I experiments, this nicked λ DNA was incubated with proteinase K enzyme for 15 mins at 56 °C to remove the nCas9 complexes bound to the DNA. The ssDNA gap was generated in situ by force-induced melting during the single-molecule experiment. The ssDNA gap in the Atto647N-labelled 17.8-kb-long dsDNA with site-specific nicks (LUMICKS) used for characterizing the spatial resolution of the single-molecule assay (Supplementary Fig. 3) was also generated in a similar manner in situ.

The DNA with a four-way junction was prepared by ligating 7.5 kb handles with phosphorylated, hairpin-forming, synthetic oligonucleotides, as described in ref. ²⁶. The handles were prepared by PCR using λ DNA as a template and two modified primers. Primer 1 contains four biotin residues at its 5′ end. Primer 2 is 5′-phosphorylated and contains a single abasic site 10 nt from the 5′ end (Supplementary Table 1). A proofreading polymerase Pfu Ultra II Fusion HS (Agilent) was used to generate a 10 nt overhang beyond the abasic site, which enables ligation of the hairpin-forming oligonucleotides. The PCR products were purified on a QIAquick PCR purification column (Qiagen). The hairpin-forming, phosphorylated oligonucleotides with 10 nt 5′ overhangs were annealed in a hybridization buffer (10 mM Tris-HCl (pH 7.5), 50 mM NaCl) that was slow-cooled from 80 °C to room temperature. The annealed product was ligated to 7.5 kb handles by incubating equimolar ratio (3 pmol) with 400 U of T4 DNA ligase (New England Biolabs) in a ligase buffer at 16 °C for 8 h. To stop the reaction, 20 mM EDTA, 0.017% SDS was added and the solution was further incubated at 65 °C for 10 min. The product was purified by electrophoresis in a 1x TAE, 0.6% agarose, 1x TAE gel followed by electroelution of the excised product band and ethanol precipitation.

Optical tweezers experiments

Single-molecule experiments were performed on a C-trap (LUMICKS) integrating optical tweezers, confocal fluorescence microscopy and microfluidics. The five-channel laminar flow cell was passivated using 0.5% (w/v) Pluronics F128 in phosphate buffered saline (PBS), and subsequently with bovine serum albumin (BSA) (1 mg ml⁻¹). Streptavidin-coated polystyrene beads, 0.005% w/v (4.8 μm or 4.35 μm, Spherotech), were injected into channel 1. Biotin-labelled DNA molecules (about 2 pM) were flowed into channel 2. Buffer A, containing 20 mM HEPES (pH 7.5), 75 mM NaCl, 0.5 mg ml⁻¹ BSA and 1 mM TCEP, was injected into channel 3. D2–I was diluted to 5 nM in buffer A and injected into channel 4. For experiments containing RPA, eGFP-RPA was diluted to 800 pM in buffer A and injected into channel 5. For ubD2–I experiments, the complex was formed by incubating purified ubFANCD2 and FANCI with relaxed, optically trapped λ DNA (less than 1 pN in the protein channel). The ubD2–I complex on DNA was then moved to the protein-free channel and kymographs were acquired at 15 pN.

The optical trap was calibrated to achieve a trap stiffness of 0.2–0.3 pN nm⁻¹. After optically trapping two beads, the DNA molecule was suspended between the beads in channel 2. The presence of the DNA tether was verified by measuring a force–extension curve in channel 3 using a constant pulling rate of 0.2 μm s⁻¹ and the acquisition rate of 60 Hz. Subsequently, the DNA tether was moved to the protein channel (channel 4) and incubated for 10–30 s. In most of the experiments, this loading step was performed using DNA held at very low force (less than 1 pN). After protein loading, the sliding experiment was carried out in the same buffer (channel 3) but without free protein. The confocal images were acquired in the absence of flow, using 488, 532 and 638 nm lasers (eGFP, Cy3 and Cy5 excitation, respectively) at a laser power of less than 3 μW. Fluorescence emission was recorded using blue (512/25 nm), green (585/75 nm) and red (640LP) filters. Kymographs were acquired by scanning the DNA contour with a pixel dwell time of 0.1–0.2 ms px⁻¹, resulting in frame rates in the range of about 100 ms, depending on the DNA tether length. All experiments were performed at room temperature (22 °C).

Single-molecule data interpretation

Raw data exported from LUMICKS Bluelake as .h5 files were processed with custom-written Jupyter Notebooks in Python 3.9 using LUMICKS Pylake v.1.2.1, numpy v.1.26.0, matplotlib v.3.7.2, scipy v.1.11.3 and peakutils v.1.3.4 (https://github.com/singlemoleculegroup). Additional adjustments (colour contrast, cropping) were performed in Fiji⁴¹. Final graphs were generated in Prism.

The particle localization and the mean square displacement (MSD) analysis were incorporated in the same Jupyter Notebook workspace using a custom tracking algorithm. The script generates a list of points that correspond with the localization of the fluorescent molecule in each time frame. Simultaneously, a list of photon counts (intensities) in each pixel in the red and green channels was generated.

MSDs from the resolved, unprocessed trajectory were calculated using the formula below:

where N is the number of frames in the kymograph, n is the size of the moving window (corresponding to the lag time τ) ranging from 1 to N − 1, X_i is the particle position at the frame i and b is the offset. The average diffusion coefficient (D) of the particle was obtained from the slope of the linear fit of MSD as a function of τ between 0.15 < τ < 1 s. To estimate the speed of the protein, the total displacement of the particle over time was calculated. Before the speed calculation, the trajectory was smoothed using a Savitzky–Golay filter.

For the rolling diffusion analysis, the resolved, unsmoothed trajectory was divided into several rolling windows (with a window size of 16 time points, corresponding to roughly 1 s). MSDs were calculated for each window (using equation (1) above). The MSDs (for the first three lag times or τs) were fit to a straight line using the relation

to yield the rolling diffusion coefficient (D_roll). Given the small number of sample points used for each window, the offset parameter used for the average D analysis was omitted for the D_roll analysis. The threshold D_roll used to distinguish between stalled and diffusing D2–I molecules was set to 6.4 × 10⁻³ μm² s⁻¹, which was thrice the average D_roll for the static nCas9 (Cy3) molecule on λ DNA (Supplementary Fig. 3).

The intensity of the trajectory at each time frame was calculated by integrating the total intensities of six pixels in the vicinity of the centre of the Gaussian peak (3 and −3). The intensity analysis from FRET quantification was done in IGOR. The intensities of donor and acceptor (Cy3- and Cy5-labelled units, respectively), I_{D_raw} and I_{A_raw}, were filtered using the smoothing BOX function over six data points.

Subsequently, fluorescence bleed-through correction was performed. The average filtered intensities of the green (I_{D_avg}) and red channels (I_{A_avg}) after acceptor photobleaching were calculated. The correction factor \({\alpha }\) was calculated as:

The corrected acceptor intensity was therefore:

FRET was calculated using the corrected acceptor intensity:

For the lifetime analysis, dwell times of D2–I stalled at the ss–dsDNA junctions were estimated directly from the kymographs. The survival probability (S) for the dwell-time distribution was calculated as

where CDF is the cumulative density function of the distribution. The survival probability was subsequently fit to a double exponential decay to the form

yielding the average lifetimes of the two kinetic phases, t₁ and t₂.

Final plots were generated using Prism, Wavemetrics IGOR 8 or in Jupyter Notebooks using matplotlib v.3.7.2.

Estimation of protein density on λ DNA

To compare the loading efficiency of D2–I on λ DNA (Extended Data Fig. 2a,b) we estimated the number of bound D2–I molecules at different NaCl concentrations (150, 100, 75 and 37 mM). We used Fiji⁴¹ to plot the two-dimensional profile along λ DNA (30 s after starting the kymograph acquisition). To estimate the number of bound D2–I molecules, we calculated the total area delimited by the distance (x axis) and the pixel intensity (y axis) and divided it by the area corresponding to a single peak (a single molecule). This process was performed for three representative kymographs for each NaCl concentration.

Electron microscopy and image processing

ss–dsDNA was prepared by incubating oligos J1 and J2 (both containing 3′-OH and 5′-OH ends; purchased from IDT; Supplementary Table 1) at 95 °C for 5 min and slowly cooling down to room temperature over 3 h. For cryo-EM analysis of chicken D2–I bound to ss–dsDNA, we incubated 1 µM D2–I complex with 3 µM ss–dsDNA in imaging buffer (used for single-molecule experiments: 20 mM HEPES (pH 7.5), 75 mM NaCl and 1 mM TCEP) for 15 min at 22 °C. A total of 3 µl of sample was applied onto plasma-cleaned Quantifoil 1.2/1.3 grids for 3.5 s before blotting and vitrification in liquid ethane using a Vitrobot Mark IV (Thermo Fisher) at 4 °C and 100% humidity. The grids were imaged using EPU v.3.4.0 on a Titan Krios (Thermo Fisher) operated at 300 keV using a Gatan K3 detector in super-resolution model at the Electron Bio-Imaging Centre (eBIC). About 20,000 videos were collected at a pixel size of 0.831 Å. We used a defocus range spanning −1.2 to −2.8 in 0.3 µm steps and a total dose of roughly 40 e/Ų per image.

All image processing was performed using RELION v.4.0 (ref. ⁴²) unless otherwise stated (Supplementary Fig. 4). Multiframe videos were drift corrected using 5 × 5 patches in MotionCorr2 (ref. ⁴³) and defocus was estimated using CTFFIND4 (ref. ⁴⁴). Particle picking was performed using crYOLO⁴⁵ in trained mode after manual picking of about 3,000 particles. Around 6 million particles were initially extracted with a pixel size of roughly 2.9 Å, with intentional overpicking to prevent missing rare views of the complex. After initial two-dimensional classification, about 2.4 million particles were subjected to three-dimensional classification, using as a reference a low-pass filtered (60 Å) map generated ab initio in RELION. Classes showing clear DNA density were selected and refined after re-extraction with the original pixel size of 0.831 Å. The defocus values of the consensus reconstruction were further refined using CTF refinement, followed by Bayesian polishing and a further round of CTF refinement. The resulting consensus refinement was subjected to three-dimensional classification without image alignment. The best classes were combined and, after refinement, postprocessed using a soft mask to correct for modulation transfer function of the detector and sharpened with a B factor of −120, as determined using RELION. Fourier shell correlation curves generated after postprocessing using a cut-off value of 0.143 yielded a map with an estimated resolution of 3.58 Å. Local resolution was calculated in RELION and displayed onto the consensus refinement in ChimeraX⁴⁶. Map sharpening of the refined maps using DeepEMhancer⁴⁷ yielded improved densities in the N termini of FANCD2 and FANCI, and aided in model building.

Linear dsDNA was prepared by incubating oligos P1 and P7 (purchased from IDT; Supplementary Table 1) at 95 °C for 5 min and slowly cooling down to room temperature over 3 h. The dsDNA–D2–I structures for both chicken and human were obtained using the same procedures in grid preparation, data acquisition and essentially the same processing strategy as described above (detailed in Supplementary Fig. 4 for chicken dsDNA–D2–I).

Structure modelling

The structure of Gallus gallus D2–I bound to dsDNA (PDB 8A2Q)¹¹ was rigidly fitted into the map of D2–I bound to ss–dsDNA using Chimera⁴⁸. After manual inspection and adjustment in Coot⁴⁰ and ISOLDE⁴⁹, the model was iteratively refined in Coot and Phenix⁵⁰. An idealized dsDNA of 24 bp was placed and refined into the duplex density using ISOLDE with distance restraints, followed by refinement in Coot and Phenix. All models and maps were inspected in ChimeraX, which was also used to prepare figures.

DT40 cell culture and mutant generation

DT40 cells (obtained from Dr J. Young at the Institute of Animal Health, Compton, Berkshire, UK) were cultured as previously described⁵¹. The DT40 cell line was authenticated by Bu-1a and b expression and immunoglobulin gene sequences. No Mycoplasma contamination was detected. The DT40 FANCD2 knockout cell line was generated by Yamamoto et al.⁵². FANCD2 point mutations were generated using Quick Change II XL Site-Directed Mutagenesis Kit (Agilent), according to the manufacturer’s instructions, in pcDNA3.1 GFP-chFANCD2 plasmid⁵². Primers for site-directed mutagenesis were designed using the Agilent QuikChange Primer Design online tool, and sequences are reported in Supplementary Table 1. Each mutant plasmid was then checked by Nanopore30 sequencing (Source Bioscience). Stable non-targeted transfections of DT40 FANCD2^−/− were performed as described in ref. ⁵³ and puromycin-resistant clones selected.

Colony survival assay

Colony survival assays were performed as described in ref. ⁵⁴ replacing D-MEM by RPMI Medium 1640 (Gibco), under penicillin/streptomycin selection, and Puromycin (1 µg ml⁻¹) for FANCD2^−/− mutants. For each mutant, two independent clones and six replicates per clone were performed. Cisplatin (Merck) was freshly prepared in a 0.9% saline solution, and cells were treated for an hour with the indicated doses. Cells were then washed, plated on methylcellulose and cultured until colonies were visible (about 2 weeks). The relative clonogenic survival, compared with untreated cells, was then calculated for each replicate. Replicates were then pooled, and statistical testing was performed by the R package CFAssay using two-way analysis of the cell survival data by fitting a linear quadratic model with maximum likelihood (P values reported in Supplementary Table 2).

Preparation of Xenopus egg extracts

The eggs of Xenopus laevis female frogs (aged more than 2 years, purchased from Nasco) were used for extract preparation. Nucleoplasmic extract (NPE) and high-speed supernatant (HSS) were prepared as previously described⁵⁵. All animal procedures were performed in accordance with national animal welfare laws and were reviewed by the Animal Ethics Committee of the Royal Netherlands Academy of Arts and Sciences (KNAW). All animal experiments were conducted under a project licence granted by the Central Committee Animal Experimentation (CCD) of the Dutch government and approved by the Hubrecht Institute Animal Welfare Body (IvD), with project licence number AVD80100202216633. Sample sizes were chosen on the basis of previous experience; randomization and blinding are not relevant to this study.

DNA repair assay in Xenopus extracts

DNA replication and preparation of Xenopus egg extracts (HSS and NPE) were performed as described previously^31,56,57. Preparation of plasmid with a site-specific cisplatin ICL (pICL) and ICL repair assays were performed as described^30,31,58. Briefly, pICL was incubated with HSS containing ³²P-a-dCTP for 20 min, following addition of two volumes of NPE (t = 0). Aliquots of replication reaction were stopped at various times with ten volumes of Stop Solution II (0.5% SDS, 10 mM EDTA and 50 mM Tris (pH 7.5)). Samples were incubated with proteinase K (0.5 μg μl⁻¹) for 1 h at 37 °C each. DNA was extracted using phenol/chloroform, ethanol-precipitated in the presence of glycogen (30 mg ml⁻¹) and resuspended in TE buffer. ICL repair was analysed by digesting 1 μl of extracted DNA with HincII, or HincII and SapI, separation on a 0.8% native agarose gel and quantification using autoradiography. Absolute ICL repair efficiency was calculated as described³¹. For reconstitutions with D2–I proteins, replication reactions were supplemented with approximately 60 nM of recombinant Xenopus laevis D2^WT–I or D2^EEEE–I or D2^EERR–I or D2^KREE–I.

Antibodies and immunodepletions

The antibody used to deplete and blot for xlFANCD2 was previously described^30,31. Histone H3 antibody was purchased from Abcam (ab1791) and used in 1:4,000 dilution. The xlFANCD2 antibody was affinity purified against the antigen used to raise the antibody. To deplete Xenopus egg extracts of FANCD2, Dynabeads Protein A beads (Thermo Fisher Scientific) were incubated with the affinity-purified antibody to their maximum binding capacity. One and a half volumes of the antibody-coated beads were then mixed with one volume of pre-cleared HSS or NPE and incubated for 30 min at room temperature. Mock depletions were performed using non-specific immunoglobulin G from rabbit serum (Sigma-Aldrich). Depleted extracts were collected and immediately used for replication assays.

Plasmid pull-down

Plasmid pull-downs were performed as previously described⁵⁹. Briefly, streptavidin-coupled magnetic beads (Dynabeads M-280, Invitrogen; 6 μl per pull-down) were washed with 50 mM Tris (pH 7.5), 150 mM NaCl, 1 mM EDTA (pH 8) and 0.02% Tween-20. Biotinylated LacR was added to the beads (4 pmol per microlitre of beads) and incubated at room temperature for 45 min. The beads were washed with 10 mM HEPES (pH 7.7), 50 mM KCl, 2.5 mM MgCl₂, 250 mM sucrose, 0.25 mg ml⁻¹ BSA and 0.02% Tween-20 and resuspended in 40 μl of the same buffer. The bead suspension was stored on ice until needed. At indicated time points, 8 μl of the replication reaction was gently mixed with LacR–streptavidin Dynabeads. The suspension was incubated at 0 °C for 30 min. The beads were washed with 10 mM HEPES (pH 7.7), 50 mM KCl, 2.5 mM MgCl₂, 0.25 mg ml⁻¹ BSA, 100 mM NaCl and 0.5% Triton X-100. All residual buffer was removed and the beads were resuspended in 2× SDS sample buffer (150 mM Tris-HCl (pH 6.8), 5% SDS, 0.05% bromophenol blue and 20% glycerol). DNA-bound proteins were then resolved by SDS–PAGE and visualized by western blotting with the indicated antibodies (anti-FANCD2 at 1:6,000; anti-Histone-H3 at 1:4,000). The recombinant D2–I complexes used for plasmid pull-down experiments were incubated with PreScission Protease (Cytiva) for 1 h at room temperature before being added to extract. This allows for the cleavage of the strep-tag from the recombinant FANCD2 to prevent direct binding of the protein to the streptavidin beads.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.